mad-0.15.0b is now three packages: libid3tag-0.15.0b.tar.gz libmad-0.15.0b.tar.gz madplay-0.15.0b.tar.gz these are separate ebuilds for those packages. madplay DEPEND's on the same version of the other two, and PROVIDE's media-sound/mad... is that an error? additionally there is a bug in alsa playback in madplay, patch provided
Created attachment 14639 [details] ebuilds for new version of mad
Created attachment 14640 [details] ebuilds for new version of mad
Created attachment 14641 [details] ebuilds for new version of mad
media-sound/mad-0.15.0b is now in portage and is a metapackage for libmad, libid3tag, and madplay. Most packages will only depend on the libraries... So ebuilds should still depend specifically on libmad and libid3tag instead of the more general media-sound/mad...
Some issues: econf, emake, and einstall should not be have the '|| die', as per the docs in 'man 5 ebuild' they each do their own failure checking should not madplay explicitly $DEPEND on libmad and libid3tag, as opposed to relying on the user installing via the metabuild? --with-gnu-ld tells configure to /assume/ a GNU ld; will it not detect its presence correctly if allowed to test it? Isn't that one of those "if it ain't broke don't fix it" things? --with-experimental in libmad should probably be avoided; probably contact R. Leslie if you really think it ought to be deployed that way to make sure it won't do anything /too/ wacky --enable-accuracy should probably be turned on in libmad; it a very efficient decoder anyway, I doubt anyone is listening to MP3's on a 386 these days. Better yet, is there some detection mechanism that "knows" whether it's running on a fast machine or on an antique? --with-esd in madplay really needs to be switched by the USE variable --with-alsa in madplay PROBABLY WON'T WORK without the patch I provided. I wrote the alsa support in madplay myself, originally against 0.14.0 if memory serves. Robert didn't test it. It didn't work on my own boxen until I located and patched some problems. IMO it's unlikely to work on anyone else's either, unless kernel 2.5.75 has major mmap bugs that i didn't bother reading about.
--with-alsa and --with-esd cannot be turned on at the same time. recommend: USE="alsa" && myconf="$myconf --with-alsa" \ || USE="esd" && myconf="$myconf --with-esd"